CIOs say data engineers, AI architects hardest roles to fill

What's the hardest IT role to hire for today?

It's always changing, but today, it's folks who have deep expertise in AI infrastructure, a lot of data engineering folks, a lot of data science and business intelligence engineers. Those folks are starting to transition into that AI world. A lot of them have made the transition, but finding someone with a lot of experience in best practices, benchmarking -- someone who's seen things at scale -- that's harder to find.

AI architects -- people who know how to fit all the pieces together and have done that at scale -- can be tricky to find. We don't do a lot of homegrown AI. For instance, we're not building proprietary models. I may have a different answer for you if we had a lot of proprietary stuff.

-- Eric Johnson, CIO, PagerDuty

CIOs get the data scientists wrong. People over-pivot on data scientists because they think they must somehow magically build models, and that data scientists are going to solve their data gap. When AI fails, it's often because the underlying data is garbage. So, data engineering is probably the hardest and most critical role to hire for right now, because bad data leads to higher hallucinatory rates, and then you get bad information and make bad decisions. If you're not working with massaged, curated data, then your AI projects will fail.

It's extremely hard to find and extremely hard to know if they actually know what they're doing. If you look at the data tools today, man, there's been an explosive growth within the last few years alone. Are they up to speed on everything? Are they going to have the mentality to keep up with the speed of change in AI and the underlying data structures?

-- Doug Gilbert, CIO and chief digital officer, Sutherland

I wouldn't put it to specifically one role, but I think AI architects. A lot of people with data analytics backgrounds think that they're AI gurus, but it's far more complex and fast-moving than that.

Most organizations are finding it difficult to work out how their ecosystem and architecture respond to infusions of AI, agents and orchestration layers -- that's a new field of architecture that no one's got a good discipline in yet. Everyone's good at the coding and creating agents, but no one's good at architecting the whole ecosystem and working out that if you have thousands of agents running around 24 hours a day, plus agents coming into your organization from third-party shippers, how do you manage, orchestrate and govern that architecture securely?

-- Joe Locandro, CIO, Rimini Street

The most difficult roles are professionals who can bridge deep technical expertise with business understanding. Demand continues to grow for AI architects, AI engineers and senior data engineers, but the common thread is their ability to translate business problems into production-ready solutions rather than simply implement technology.

Developing internal talent is becoming a competitive advantage. Technical skills can be taught, but institutional knowledge, business context and trusted relationships are much harder to acquire. Organizations that invest in reskilling strong technologists will often move faster than those competing for a very limited pool of experienced AI talent.

-- Chris Campbell, CIO, DeVry University

The hardest hires today are technology teammates and leaders who can bridge business strategy and technical execution. Organizations need people who understand AI, data, cybersecurity, engineering and operations, but can also translate those capabilities into measurable business results that can be consumed by business teammates and leaders.

-- Steve Bronson, CIO, Southern Glazer's Beverage Co.

It's less about specific job titles and more about finding the right combination of skills. 

The first is what I'd call an AI-fluent technologist. We need people who understand machine learning and large language models, but who also know how to build scalable enterprise systems. Plenty of people understand AI, and plenty know software development, but finding professionals who can bring those disciplines together with strong architectural and cybersecurity principles is much harder. That's exactly the skill set required to scale AI responsibly in a highly regulated industry like claims. 

The second is the ability to bridge modern AI technologies with legacy enterprise environments. Most large organizations, including Sedgwick, operate a mix of cloud-native platforms alongside systems that have supported the business for years. The challenge is finding developers who are comfortable in both worlds and can make thoughtful architectural decisions across that environment. As more talent enters the market with primarily cloud-native experience, that hybrid expertise is becoming increasingly difficult to find. 

-- Sean Safieh, CIO of global platforms and digital solutions, Sedgwick

The hardest IT role to hire for today is a transformation leader. Five, 10, 15 years ago, you could just hire somebody who was hardcore, who could connect your systems and data, understand architecture and scale, and put it together. Today, that's not how the world works. You really need someone who understands technology, data and how data is interconnected and meaningful.

You must hire somebody who understands architecture, who is curious about the business and has some empathy. It's a big job. You're also typically still running whatever operational function you have while you're transforming that function. So, you are flying the plane, rebuilding the plane and serving a six-course dinner on the plane.

-- Kellie Romack, chief digital information officer, ServiceNow

Tim Murphy is a site editor and writer for the IT Strategy team at TechTarget.